The Size of Thoughts

by Nicholson Baker

  (1991)

  Ice Storm

  I grew up in Rochester, so I should probably know what an ice storm is, but what has just happened here is brand-new in my experience. On the night of March 3, a Sunday, some sort of strange, gentle, superchilled rain came down over a large part of western New York. It coated every power line with a perfect cylinder of ice nearly an inch thick, from which ideally spaced icicles, like the tines on a soil rake, descended, all exactly the same length. The freeze held through early Tuesday. On Monday, if you looked out any window for a few minutes, you were certain to see, against a background of glittering Ace combs, the bough of a tree come crashing down. There was no wind, nor had there been any the first night. It seemed more a demonstration of the patient principles of candlemaking than a storm. At a distance, the ice effects were white, but when you drew close enough to a tree to be surrounded by the continual worry-bead crackling of its fretwork, and stood there, ready to duck at any moment, you saw that it had become incorporated into a clear and disturbingly clinical arrangement of pristine pipettes and test tubes, each holding a once-natural element of the organism—a bud, a twig, one of those perky citizens you had been counting on to function as usual in a few months—in an elaborate cryopharmaceutical experiment.

  I drove down the streets today (Tuesday) feeling at times that it was all very familiar, that Ansel Adams calendars had prepared me for this, but then, jumping out again and again from the arty, grainy black-and-white photography that slowly moved past was a sudden apricot-colored splash of discomfort where a bough had torn itself free or a trunk had split in half. On the streets I’ve seen, half of the good big trees are ravaged. The younger ones, planted about twenty years ago to replace all the elms, are especially painful broken sights. Conifers did somewhat better than deciduous trees. The tall weeping spruce next to our house weeps more than usual but has lost no limbs. Stuck high up in it before the storm was a plastic dragon kite, which we had repeatedly tried and failed to extricate; the morning after the storm, it lay in two pieces on the grass. My wife said, “Well, at least something good has come out of this.” I said yes, it was like bombing all of Iraq to get rid of Saddam Hussein. She reminded me of Frost’s poem about whether the world will end in fire or in ice. And if we hadn’t just flown more than a hundred thousand sorties over a distant place, I would give in more to grief about all these trees, but in the face of that devastation this sort of rare and unmalicious natural catastrophe, in which nobody dies, and some leftovers spoil in some refrigerators, and people go out on tentative camera expeditions to pass the time until the cable TV comes on again, makes me think that we over here have gotten off very easily. We deserve at least this much ice after that much fire. Many of the trees will grow back, after all, as from a bad pruning. As they thaw now, the water is hearteningly visible, hurrying along the bark underneath the ice layer, like blood.

  (1991)

  Discards

  One sunny afternoon in October of 1985, a crowd of librarians and library administrators gathered near the venerable card catalog of the Health Sciences Library of the University of Maryland at Baltimore. It was time for a little celebration. Hundreds of red and blue balloons gently exerted themselves against the acoustical-tiled ceiling. From each balloon hung a piece of string; at the end of each string, tied by its guide hole, dangled a card that had been plucked at random from the library’s catalog. On the back of every card was a stamped message: GENUINE ARTIFACT FROM THE CARD CATALOG OF THE HEALTH SCIENCES LIBRARY, UNIVERSITY OF MARYLAND AT BALTIMORE. A TV crew was there, along with a few local dignitaries and a reporter from the Baltimore Sun. The group stood among the hovering cards and listened to some speeches about what a landmark day it was for the library, and about the many decisive advantages of online library catalogs—remote access, a more efficient flow of information, reduced costs, lives saved (for this was a medical institution)—and then Chancellor Edward N. Brandt, Jr., wearing a red T-shirt that said “The Great Discard,” chose a drawer of the catalog and pulled it from the cabinet. With the help of a beaming Cyril Feng, who was then the director of the library, he drew the retaining rod from the chosen drawer and let its several hundred cards ceremonially spill into a trash can decorated with colored paper.

  When the applause died down, the onlookers (there were about a hundred of them, including, in addition to librarians, a number of nursing students invited to swell the crowd) gathered as many of the balloons as they could and made their way out to a grassy plaza across the street. Reassembled, at the count of three, they released their balloons. The sunlit bibliographic payloads twirled picturesquely as they rose. Soon the wind caught them, and they disappeared off to the northeast. Some days later, one of the cards was returned to the library from a place across the Chesapeake Bay by some resident who thought, foolishly, that it was of value. Another card floated all the way to Connecticut.

  Since then, as universities and public libraries have completed the “retrospective conversions” of their catalogs to computer databases (frequently with the help of federal Title II-C money, as part of the “Strengthening Research Library Resources” program), hundreds of card catalogs, tens of millions of individual cards—cards typed on manual typewriters and early electric models; cards printed by the Library of Congress, Baker & Taylor, and OCLC; cards whose subject headings were erased with special power erasers, resembling soldering irons, and overtyped in red; cards that have been multiply revised, copied on early models of the Xerox copier, corrected in pencil, color-coded, sleeved in plastic; cards that were handwritten at the turn of the century; cards that were interfiled by generations of staffers, their edges softened by innumerable inquiring patrons—have been partially or completely destroyed, many in the past year.

  The main subject card catalog of the University of California at Berkeley was thrown out in the summer of 1993 to make way for eight study tables. Bryn Mawr’s main catalog is gone. The University of Hawaii’s main catalog is just about gone. A recycling firm called Earthworm, Inc., carted off the bulk of MIT’s cards in 1989—these now mingle unintelligibly as shoe or cereal boxes, or constitute a kind of electrical insulation known as “creping tissue.” The cards for the Math, Africana, Engineering, and Physical Sciences collections at Cornell’s John M. Olin Library are gone, with others to follow. The subject catalogs at Dartmouth, Kent State, and Boston University are gone. Harvard’s cards are going. Cards representing classical literature and philosophy have been pulled from the University of Chicago’s main catalog; some of the American-literature cards are being tossed out now.

  The New York Public Library, ahead of the game, renovated the entire ten-million-card catalog of its Research Libraries between 1977 and 1980, microfilmed it, and threw it out. Stanford is discussing when and how to recycle its catalogs. The cards dating from 1911 to 1975 at the New York State Library in Albany (where Melvil Dewey was librarian from 1889 to 1906) were thrown away last month as a consequence of a historical-preservation project involving the building in which they were stored. The catalog for the Hawthorne-Longfellow Library at Bowdoin College is 95 percent gone. The cards from UCLA’s main library are nearly gone. My own college, Haverford, stored its card catalog in an attic for two years after finishing its retrospective conversion, in 1991. But when I talked to the director of the library in August 1993, the catalog (including many thousands of handwritten cards) was being thrown out, and the cabinets were down at the physical plant being dismantled and scrapped.

  One of the odder features of this national paroxysm of shortsightedness and anti-intellectualism (“In a class with the burning of the library at Alexandria,” Helen Rand Parish, a historian specializing in the sixteenth century, said to me) is that it isn’t the result of wicked forces outside the library walls. We can’t blame Saracen sackers, B-52s, anarchists, or thieves; nor can we blame propagandistic politicians intent on revising the past, moralistic book banners, or over-acidic formulations of paper. The villains, instead, are smart, well-m
eaning library administrators, quite certain that they are only doing what is right for their institutions.

  And, incredibly, nobody is making an audible fuss about what they are up to. Nobody is grieving. On the contrary, there are balloons and nursing students; there are festive pictures in trade magazines and industry newsletters of smiling department heads wearing aprons as they dump trays of cards into rolling trash carts. At Cosumnes River College, in California, the card catalog was ceremonially put out of its misery by an official who pointed a gun at it and “shot” it. Dickinson College held a mock wake: veils were worn, hymns were sung, and the doomed catalog was decked with wax flowers and black garbage-bag bunting. “We are only too happy to throw the cards away when we’re finished,” Judith Brugger, Cornell’s catalog management and authorities librarian, told me. As an added publicity flourish during the Great Discard celebration, the Maryland Health Sciences Library published a commemorative chapbook called 101 Uses for a Dead Catalog Card. Some of the suggested uses: tablecloth-crumb scrapers, space-shuttle tiles, garden compost, fish scalers, cat litter, jousting targets, and serial way-markers for spelunkers. Use No. 100 was “Make a bonfire out of them to celebrate the end of the card catalog.”

  Is this glee really justified? Is it seemly? I called up Dale Flecker, associate director for planning and systems for the Harvard University Library, and asked him if he felt any regret at the throwing out of card catalogs. “In general not,” he said. “They were a wonderful idea for their time. Their time was about a century, and it’s now coming to a close.” Mr. Flecker is currently overseeing what is perhaps the largest single retrospective-conversion project ever attempted: the transfer of the information on five million pre-1980 cards, taken from approximately a hundred libraries in the university—including millions from Widener, the main library—to Harvard’s online catalog, HOLLIS. The conversion of the Widener catalog is in a fairly early stage, and it is in many ways a model of planning and forethought. Each of Widener’s cards was microfilmed before the catalog was “disassembled,” using a custom-made card-feeding machine. (Microfilming is, of course, a luxury few libraries can afford. Half the cost of Harvard’s campus-wide conversion project is being funded with federal grants and large private gifts.) An image of the front of every card for Widener thus now exists on microfiche, available to users in a room off the lobby. (Any information on the backs of the cards—and many notes do carry over—was not photographed.) Every Tuesday, Harvard sends batches of cards, twenty thousand or so at a time, via UPS, to a nonprofit corporation in Dublin, Ohio (a suburb of Columbus), called OCLC.

  These initials once stood for Ohio College Library Center, but now—since the company has grown over the past twenty-three years into an international eminence in the information industry, with yearly tax-free revenues of close to a hundred million dollars—they stand for Online Computer Library Center. OCLC owns the largest database of bibliographic information in the world, and it offers a service called RETROCON, contracting with libraries to transfer old catalog cards to “machine-readable form,” at anywhere from fifty cents to six dollars per card.

  The RETROCON business is good right now. When I visited OCLC in September 1993, there were three shifts of sixty operators each, plowing through RETROCONs for about forty different libraries, including, besides Harvard, the Los Angeles Public Library, a consortium of French libraries, Hughes Aircraft, Brown, Queen’s University of Belfast, Northwestern, the library at Kew Gardens, the Cincinnati Public Library, the San Francisco Public Library, and Waseda University in Japan. Each operator, though initially hired as a temp, and, in some cases, no more than high-school-educated, is the survivor of a rigorous two-week training program. (Most hires do not make it through training.) At first their work is reviewed every day by a supervisor, and subsequently it is spot-checked for accuracy at unexpected times. “Our standards are very, very high,” OCLC’s Maureen Finn told me as we stood next to shelves holding hundreds of long gray cardboard boxes that said “Widener.” “Libraries are entrusting us with the history of their library collection. We have to make sure that we’ve got good people who are going to do the work correctly, and that we can trust them.”

  Maureen Finn runs RETROCON and, from what I saw, she runs it exceedingly well. Some of her people have been converting catalog cards, with apparent contentment, for ten years or more. It is true that they seldom have formal training in the intricacies of the cataloger’s art: they are not necessarily up on the elaborate Anglo-American cataloging rules, nor are they current concerning the periodic RI’s, or Rule Interpretations, issued by the Library of Congress, which—like hermeneutical dispatches from the IRS or the Financial Accounting Standards Board—adjudicate perplexing cases as they emerge. Most operators have had no library-school courses in the Dewey decimal system (still thriving, by the way, and very big in Europe), or the Sears List of Subject Headings, or Cutter numbers, or the abbreviational niceties of the International Standard Bibliographic Description (ISBD) format—but they do apparently pick up a great deal as they go. “I would put any of these people up against an MLS cataloger any day,” Ms. Finn said to me—“MLS” meaning Master of Library Science. “I think that their breadth of knowledge is quite extensive.”

  Well, it better be. The next card that a second-shift operator props onto the top of the keyboard, at ten-fifteen at night, could be in any of three hundred roman-type or transliterated languages, about anything at all, covering any period of human or interstellar history. His or her job is to find a match, or “hit,” for that card in OCLC’s huge database. A hit brings up a specially formatted computer record called a MARC record (MARC stands for MAchine Readable Cataloging)—a daunting set of numbered fields and odd symbols developed by the Library of Congress in the sixties, redolent of unfriendly first-generation database interfaces—describing the particular edition of a book (or map, or videotape) which corresponds to the one that the physical card stands for. The hit rate varies from project to project, but it is running at around 70 percent for Harvard.

  If the operator does find a matching record, she (there are slightly more women than men) then must make modifications to it onscreen to suit the source library’s idiosyncrasies of call numbering, entering as much or as little of the card’s supplemental information as is required by the contract between OCLC and that particular library. (The more modification a library demands of each MARC record, the more it costs.) In Harvard’s case she typically accepts the record as is, even when the original card bears additional subject headings or enriching notes of various kinds.

  If her search through the database does not turn up a match for the card, she must enter the information on it from scratch. This is, or should be, a complex undertaking, demanding judgment as well as accuracy, since not only does the card’s content need to be transferred perfectly, without reference to the book itself, in a way that will accord with records for other books by that writer in that library, but the necessary “access points” have to be tagged correctly with numeric codes—access points being those fields in the record that a library user will be able to search for, such as title, author, corporate author, and subject. If someone has a bad night and creates a flawed access point for the title of a book, library patrons will simply not find the book in the library’s online catalog when they search for it by title.

  We shouldn’t leap to the conclusion that just because the RETROCON staff is composed of temps and ex-temps, many without BAs, it will be doing awful things to Harvard’s cards. I can vouch for the devotedness and hard work of (some) temps, having temped for several years myself. And airline pilots, some of whom never went to college either, land jumbo jets a thousand times a day with an exceedingly low error rate. Still, we have to wonder whether the Harvard community is expecting something of Maureen Finn’s staff at RETROCON that the staff can’t possibly deliver—something, in fact, that no outside group of clerical workers, sitting hundreds of miles away from the books in question, no matter how
well trained and closely monitored they may be, can deliver. Even OCLC, one of the very best retrospective-conversion contractors in the business, is bound to make thousands of typos in the course of a huge project like the Harvard “recon.” There is of necessity going to be a layer of error introduced into Harvard’s online catalog. (The official error rate is “less than one per cent”—which, for five million cards, is less than fifty thousand records.) And some of the mistakes, though tiny in relation to the extraordinary size of the database, will—in the same way that errors in TRWs database can do vexing things to one’s credit rating—be very significant for individual scholars in quest of a particular book. Nor will there be a wave of compensating corrections as a result of the conversion, since the books themselves aren’t there for a zealous operator to check against.